Skip to main content

Introduction to Data Science Using R Runestone Version 1

Dr. Jan Pearce, Editor

Section 12.5 Test Yourself

Checkpoint 12.5.1.

What is a kernel in the context of Data Science and Support Vector Machines?
🔗
  • A subset of the data set aside to guage the accuracy and bias of training data.
    🔗
  • Incorrect. This data is called testing data.
  • The core part of an operating system that manages system resources and allows software to interact with hardware.
    🔗
  • While technically correct, this particular kernal is more of a focus in Operating Systems courses. The definition of kernal in this chapter relates to Data Science and SVMs.
  • A function that separates data into two or more labels.
    🔗
  • Incorrect. This is the definition of a separation function.
  • A mapping function that transforms data into a higher-dimensional space.
    🔗
  • Correct!
🔗

Checkpoint 12.5.2.

Why should data be randomized before being divided into training data and test data?
🔗
  • Most randomize functions also clean data.
    🔗
  • Incorrect. It is best practice to create functions that do one thing and one thing only.
  • The data may be sorted or grouped in some way. This can lead to training data bias.
    🔗
  • Correct!
  • Randomizing data increases the speed at which the data is processed.
    🔗
  • Incorrect. Some sorting algorithms have a better time complexity with randomized data, but this isn’t the case with dividing data.
  • There is no need to randomize data before dividing it into training and testing data.
    🔗
  • Incorrect. Unless it can be confirmed that there is absolutely no patterns or ordering in the data, randomizing the data is necessary.
🔗
Chapter Challenge
🔗
Look up the term "confusion matrix" and then follow-up on some other terms such as Type I error, Type II error, sensitivity, and specificity. Think about how the support vector machine model could be modified to do better at either sensitivity or specificity.
🔗
For a super challenge, try using another dataset with the kernlab svm technology. There is a dataset called promotergene that is built into the kernlab package. You could also load up your own data set and try creating an svm model from that.
🔗
You have attempted of activities on this page.
🔗
PrevTopNext